Send Orders for Reprints to
[email protected] 16
The Open Neuroimaging Journal, 2018, 12, 16-29
The Open Neuroimaging Journal Content list available at: www.benthamopen.com/TONIJ/ DOI: 10.2174/1874440001812010016
RESEARCH ARTICLE
Brain Responses Underlying Anthropomorphism, Agency, and Social Attribution in Autism Spectrum Disorder Carla J. Ammons, Constance F. Doss, David Bala and Rajesh K. Kana* Department of Psychology, University of Alabama at Birmingham, Birmingham, AL, USA Received: November 28, 2017
Revised: February 27, 2018
Accepted: March 03, 2018
Abstract: Background: Theory of Mind (ToM), the ability to attribute mental states to oneself and others, is frequently impaired in Autism Spectrum Disorder (ASD) and may result from altered activation of social brain regions. Conversely, Typically Developing (TD) individuals overextend ToM and show a strong tendency to anthropomorphize and interpret biological motion in the environment. Less is known about how the degree of anthropomorphism influences intentional attribution and engagement of the social brain in ASD. Objective: This fMRI study examines the extent of anthropomorphism, its role in social attribution, and the underlying neural responses in ASD and TD using a series of human stick figures and geometrical shapes. Methods: 14 ASD and 14 TD adults watched videos of stick figures and triangles interacting in random or socially meaningful ways while in an fMRI scanner. In addition, they completed out-of-scanner measures of ToM skill and real-world social deficits. Whole brain statistical analysis was performed for regression and within and between group comparisons of all conditions using SPM12’s implementation of the general linear model. Results: ToM network regions were activated in response to social movement and human-like characters in ASD and TD. In addition, greater ToM ability was associated with increased TPJ and MPFC activity while watching stick figures; whereas more severe social symptoms were associated with reduced right TPJ activation in response to social movement. Conclusion: These results suggest that degree of anthropomorphism does not differentially affect social attribution in ASD and highlights the importance of TPJ in ToM and social attribution. Keywords: Autism, Animation, Anthropomorphism, Theory of Mind, Social attribution, Functional MRI, Social brain.
1. INTRODUCTION The ability to understand and predict the behavior of other agents, who possess their own thoughts, desires, goals, and intentions, agents whose motives are often hidden and must be inferred from their actions, is one of the most complex cognitive tasks the human brain has to accomplish [1]. It is also one with enormous evolutionary benefit [2, 3]. Rapidly detecting the movements of agents in the social world allows for speedy threat detection and accurate prediction of behavior. This ability relies on two processes, identification of biological movement in the environment * Address correspondence to this authors at the Department of Psychology, University of Alabama, Birmingham, CIRC 235G, 1719 6th Ave South, Birmingham, AL 35294-0021, USA, Tel: (205) 934-3171; Fax: (205) 975-6330, Email:
[email protected]
1874-4400/18
2018 Bentham Open
Brain Responses Underlying Anthropomorphism
The Open Neuroimaging Journal, 2018, Volume 12 17
and attribution of intentionality to the entity engaged in that movement. Both processes impact adaptive social behavior including day-to-day social interaction and nonverbal communication [4]. In Social Psychology, the term Theory of Mind (ToM) is used to describe the ability to attribute mental states to oneself and others. The ToM account of Autism Spectrum Disorder (ASD) attributes the social communication deficits in ASD to a dysfunctional ToM system [5, 6]. While this model may not explain the entire spectrum of autism symptoms, deficits in ToM have been consistently documented and linked to real-life social difficulties [6, 7]. The development of basic ToM skills in children with ASD is often delayed compared to mental age-matched peers and to the development of their other cognitive skills [8, 9]. Furthermore, precursors to ToM including joint-attention, gaze following, and pretend play are often absent, limited, or delayed in infants later diagnosed with autism [10 - 13]. High functioning adolescents and adults with ASD, despite demonstrating substantial real world social difficulties, are often able to pass commonly used laboratory tasks of ToM [14 - 16]. It is noteworthy that performance on traditional ToM tasks correlates with verbal IQ in typically developing and patient populations [8, 17]. Verbal reasoning, therefore, may help inflate performance. Another aspect of traditional ToM tasks is the direct cueing of attention to others’ thoughts, intentions, or emotions since many tasks explicitly ask what someone is thinking or highlight relevant social cues in context [15]. This explicit cueing may provide an advantage that is not present in real-world social settings, thus not capturing the true social difficulties of the higher functioning population. In order to address the limitations of traditional ToM tasks, the social attribution task was adapted from a classic study demonstrating that the movement of simple geometric shapes gives rise to an impression of intentional action and provokes the attribution of complex mental states [18]. In this task, shapes may be perceived as biological agents with beliefs, desires, and intention. Typically developing (TD) individuals identify the social nature of the movement relatively quickly and provide descriptions that include social motives (i.e. protecting, bullying), relationships between the characters (i.e. mother and child, friends), and attribution of mental states (i.e. scared, shy). In contrast, individuals with ASD, while accurately describing mechanical or goal directed movements of the shapes, did not pick up on the social or relational aspects of the shape movement or tend to use mental state language to describe the action despite being able to pass other higher order ToM tasks [19 - 21]. This suggests that even in high functioning individuals with ASD, spontaneous attribution of intentions to events in the social world may be problematic and the social attribution task is sensitive to subtle social cognition deficits. Neuroimaging studies across different ToM tasks have identified a common network of brain regions that show increased activity. This includes the Medial Prefrontal Cortex (MPFC), the posterior superior temporal sulcus (pSTS) at the Temporal-Parietal Junction (TPJ), and the temporal pole anterior to the amygdala [22 - 24]. These regions are consistently activated when making inferences about others’ mental states, regardless of the specific task demands (i.e. false belief tasks, verbally-based mentalizing tasks, picture completion or comic strip vignettes, and animated shape paradigms) suggesting an underlying, integrated network. Studies in ASD have shown hypoactivation of these ToM regions, even in the presence of similar behavioral performance [25 - 27]. Interestingly, many of these same regions are involved in the detection of biological motion, particularly the right STS [28, 29]. It has been proposed that biological motion processing is closely associated with the development of social cognitive abilities [29] and individuals with ASD have also demonstrated deficits in biological motion processing [30 - 32]. This includes difficulty detecting camouflaged biological motion [33] and discriminating biological motion in point light displays [34]. Three previous studies have examined the neural responses underlying the attribution of mental states to moving shapes in ASD, a task that combines ToM, anthropomorphism, and biological motion processing. In Castelli et al.’s [25] study, high functioning adults with ASD activated the MPFC, STS, and temporal pole areas to a lesser extent than well-matched control participants when viewing shapes moving interactively with social motives. Furthermore, extrastriate regions which activated strongly during the mentalizing condition for both groups showed reduced connectivity to the STS in the ASD group [25] suggesting reduced communication between these regions. In two studies from Kana and colleagues, high functioning adults with ASD demonstrated reduced activation in ToM and mirror network regions, including MPFC, TPJ, and cingulate cortex, when viewing social animations as well as reduced connectivity between frontal and posterior ToM regions [26, 27]. Furthermore, activation in the right TPJ was correlated with ToM ability in the ASD group [27]. These studies suggest that regions of the ToM network are underactive and underconnected in ASD when making spontaneous social attributions. What is less understood is how specific characteristics of the agents performing the action influence social attribution and anthropomorphism in ASD. The tendency toward anthropomorphism is well-documented in TD individuals; where attributing human-like beliefs, intentions, and emotions to animals and even inanimate objects is
18 The Open Neuroimaging Journal, 2018, Volume 12
Ammons et al.
common. Motivational aspects as well as cognitive mechanisms have been proposed to explain anthropomorphic tendency with some considering anthropomorphism the brain’s “default” processing style in the face of ambiguous input [35]. Other cognitive mechanisms including attribution of agency, biological motion detection, motor simulations and the mirror neuron system, embodied cognition, empathy, and causal reasoning have all been proposed as contributors toward anthropomorphic bias [35]. Many of these processes have also been implicated in social cognition difficulties in ASD and are building blocks in the social attribution process. However, the nature and characteristics of the agent of action can influence anthropomorphism as a response bias. For example, Chaminade et al. [36] showed that motion performed by more anthropomorphic computer generated characters reduces bias toward interpreting movement as biological, the “uncanny valley” hypothesis. At the same time, response bias towards perceiving motion as biological, regardless of character, was associated with increased activation of left STS, left TPJ, right Superior Temporal Gyrus (STG), Anterior Cingulate Cortex (ACC), and precuneus – regions considered to be crucial in processing ToM [36]. A follow-up study showed that children with ASD did not display the characteristic anthropomorphic bias of TD children [37]. Instead, children with ASD were not influenced by the degree of congruence between anthropomorphism of the character and movement suggesting they failed to integrate context and/or recognize the social information in the character. Interestingly, anthropomorphism bias has not been studied in the context of the social attribution task in ASD. The present fMRI study examines the extent of anthropomorphism, its role in social attribution, and the underlying neural responses in ASD and TD using a series of stick figure human characters and geometrical shapes. We investigate the impact of an inherently social character (a stick figure human) on perceived biological motion, social attribution, and brain activity underlying both. We presented short video clips of moving two-dimensional characters and asked participants to identify whether or not the characters were moving in an intentional, socially directed manner. Importantly, we manipulated aspects of the agent’s movements (social vs. nonsocial) as well as aspects of the agents themselves (social vs nonsocial) in order to test the differential contributions of motion processing and character features (geometrical shapes vs. humanlike stick figures) to agency and social attribution. This study is novel in that it examined both movement and agent characteristics affecting social attribution in the same well-defined sample of adults with ASD compared to age-and-IQ-matched peers. Previous research has focused almost exclusively on the movement of the characters rather than the characters themselves. As such, this study is unique in that it can comment on the impact of social aspects of the characters as well as their movement and how this differs between ASD and TD participants. Based on previous studies, we hypothesized that participants with ASD would have more difficulty detecting social motives in the movements of both humanlike and shape characters and that this difficulty would be reflected in reduced activation of key ToM regions (MPFC, pSTS, TPJ) during social attribution regardless of character. We also hypothesized that viewing more anthropomorphic characters, regardless of the type of movement they were performing, would more strongly activate social brain regions and that this would be most apparent in the TD group due to the inherent social/biological cue of the stick figure. Lastly, we expected that ToM network activity during social attribution based on movement or character information would predict general ToM abilities measured outside the scanner. 2. METHODS 2.1. Participants Functional MRI data were collected from 17 TD young adults and 17 high functioning adults with ASD. Data from three individuals in the ASD group were excluded due to equipment failure (1) and inattentive responding (2), defined as an in-window response to less than 50% of stimuli or accuracy below 65%. Three individuals from the TD group were excluded due to excessive head motion (1), equipment failure (1), and inattentive responding (1). This resulted in a final sample of 14 TD and 14 ASD participants (Table 1). Groups were matched on age, IQ, and handedness. Average head motion, defined as the root mean square of displacement, did not differ between ASD and TD groups (t(28,2)=1.47, p=0.15) and all participants retained greater than 80% of functional volumes after censoring. Participants with ASD were recruited from local service providers in the Birmingham and Tuscaloosa areas including the UAB Civitan-Sparks clinic, Mitchell’s Place, the Glenwood Foundation, and the University of Alabama ASD Clinic, as well as from previous involvement with the Cognition, Brain, and Autism Lab at UAB. Previous ASD diagnosis via the Autism Diagnostic Observation Schedule (ADOS) and/or the Autism Diagnostic Interview (ADI) was required and verified by patient records. TD participants were recruited from the UAB campus and the Birmingham area at large. All participants were screened for and excluded based on claustrophobia; body mass index exceeding 34;
Brain Responses Underlying Anthropomorphism
The Open Neuroimaging Journal, 2018, Volume 12 19
history of working with metal or non-removal metallic body implants; medical conditions affecting blood flow including diabetes, hypertension, sickle cell disease, and anemia; current use of psychotropic medications; and any reported history of neurological disorder. Prior to scanning, all study procedures were described to participants and informed consent was obtained. All aspects of this study were approved by the UAB Institutional Review Board. Table 1. Demographic information for the study sample. Values are presented as mean (standard deviation; range). T statistics and p values are for independent sample t-tests between group means. Groups ASD (n=14)
TD (n=14)
t
p
Gender
10 M; 4 F
13 M; 1 F
-
-
Handedness
13 R; 1 L
13 R; 1 Ambidextrous
-
-
Age (years)
25.07 (±6.16; 17-40)
24.86 (±5.35; 19-36)
0.10
0.92
Full Scale IQ (WASI)
117.57 (±12.97; 99-140)
117.23 (±8.78; 105-134)
0.08
0.94
Verbal IQ
117.07 (±11.16; 101-135)
112.38 (±11.12; 92-129)
1.09
0.29
Performance IQ
114.00 (±15.81; 89-138)
118.54 (±8.17; 103-132)
0.93
0.36
RAADS-R total
131.29 (±32.19; 72-181)
42.71 (±23.57; 3-78)
8.31
